Let's consider, instead of an external reference to a cohort id, that
instead selecting a cohort embeds the cohort, and its associated
conceptsets into the 'estimation' project, so that its fully
self-contained. (akin to how conceptsets get embedded into cohort
definitions). then, when we write the R code, instead of referencing
cohortid, we can instantiate the SQL for the cohort definitions, place in a
temp table, and reference them in the rest of the package, like how we've
done other prior studies. this will make the package transportable across
instances.
anthonysena
commented
6 years ago
Related to #261. Per @pbr6cornell